Here is a quick (and largely untested, apart from the toy example) hack that
should read and write UTF-8 encoded CSV files on Window. It is *not* the best
solution, which should use proper C-level code. But it could help a bit in your
case.
Best,
Philippe Grosjean
read.tableUTF8 <- function (file, ...)
{
if (l10n_info()$`UTF-8`) {
read.table(file = file, fileEncoding = "UTF-8", ...)
} else {
res <- read.table(file = file, ...) # Read in default encoding
## For each character variable, change encoding to 'UTF-8'
## For each factor, change encoding to 'UTF-8'
as.data.frame(lapply(res, function (x) switch(data.class(x),
character = {Encoding(x) <- "UTF-8"; x},
factor = {Encoding(levels(x)) <- "UTF-8"; x},
x))
)
}
}
write.tableUTF8 <- function (x, file = "", ...)
{
if (l10n_info()$`UTF-8`) {
write.table(x = x, file = file, fileEncoding = "UTF-8", ...)
} else {
## Change encoding to "bytes" and save it like that
x <- lapply(x, function (x) {
if (is.character(x)) {
Encoding(x) <- "bytes"
} else if (is.factor(x)) {
Encoding(levels(x)) <- "bytes"
}
x
})
write.table(x = x, file = file, ...)
}
}
fact <- factor(c("\u0444", "\u220F", "\u2030"))
char <- c("\u2202x", "\u2202y", "\u2202z")
dfr <- data.frame(x = 1:3, f = fact, s = I(char))
dfr
write.tableUTF8(dfr, file = "testUTF8.txt")
dfr2 <- read.tableUTF8("testUTF8.txt")
dfr2$s <- I(as.character(dfr2$s))
dfr2
identical(dfr$f, dfr2$f)
identical(dfr$s, dfr2$s)
______________________________________________
R-devel@r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/r-devel
